Chitosan-conjugated polyethersulfone membranes with epitope-imprinted polymers for hepatitis A virus adsorption: Filtration optimization, kinetic modeling, and structural characterization

Chitosan-conjugated polyethersulfone (PES) membranes functionalized with epitope-imprinted polymers (eIPs) were developed and optimized for the selective elimination of hepatitis A virus (HAV) from water. The effects of nano-molecularly imprinted polymer (nanoMIP) concentration (400–600 μg mL -1 ), incubation time (IT; 5–65 min), and filtration time (FT; 0.3–1.9 min) on permeation flux, recovered volume, flux decline ratio (FDR), HAV removal, and recovery were evaluated using response surface methodology-central composite rotatable design (RSM-CCRD). The second-order polynomial models showed high predictive accuracy (R 2 = 0.9180–0.9996). Optimized conditions (529.59 μg mL -1 nanoMIP, 20 min IT, 0.721 min FT) yielded high-reliable predicted values of 476.10 L m -2 h -1 (permeation flux), 100.65 % (recovered volume), 8.64 % (FDR), 99.94 % (HAV removal), and 33.16 % (HAV recovery). Scanning electron microscopy (SEM) and fluorescent microscopy confirmed time-dependent nanoMIP and HAV accumulation on membrane surfaces and pores. Smartphone image analysis revealed reduced grayscale brightness (230.9 to 61.44) and increased texture irregularity. SEM showed porosity reduction (4.95 % to 0.37 %) and tortuosity increase (4.49 to 16.31). Fourier-transform infrared (FTIR) spectroscopy indicated stronger C=O, N–H, and S=O interactions at higher nanoMIP loads and FT. X-Ray photoelectron spectroscopy (XPS) confirmed higher O 1s and lower C 1s intensities, indicating progressive surface oxygenation at longer FTs. Kinetic analysis showed HAV adsorption followed the fractional power model (R 2 = 0.9289, RMSE = 0.00370), suggesting diffusion-limited, multilayer binding. The integration of tailored imprinting, hydrophilicity enhancement, and process optimization enabled effective HAV capture with reduced fouling, demonstrating strong potential for virus-selective membrane filtration. • Peak HAV removal efficiency at 530 μg mL -1 nanoMIP, 20 min incubation, and 0.72 min filtration. • Increasing nanoMIP level (≤530 μg mL -1 ) improved HAV retention and flux efficiency. • Quadratic RSM-CCRD models (R 2 = 0.918–0.999) accurately predicted filtration performance. • 99.8% HAV removal and 36.6% recovery with chitosan-functionalized PES membranes. • HAV adsorption kinetics best fitted the fractional power model (R 2 = 0.929; RMSE = 0.0037).